Unverified50% confidenceBenchmarkExact time
The paper evaluated 8 LLMs over 1,292 dialogue turns and found that models with similar overall scores differ enormously in compliance rate on individual rules.
1
Sources
50%
Confidence
Long-term
Relevance
10/20/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThat experiment reached the adaptation criterion specified in the paper (three consecutive sessions passing) at session 14, with both the bold rate and list rate metrics showing a downward trend across the first 36 sessions.71% similarUnverifiedIn a test covering 8 local models and more than 5000 calls, the wording style of the prompt (plain/normal/florid) had almost no effect on accuracy on non-math tasks, with differences of no more than 4 percentage points for models of 8B and above.70% similarUnverified早期研究发现同一问题用不同表述方式提问，模型性能差异可高达数十个百分点70% similarUnverifiedIn the first round of experiments, features with verified capabilities averaged 1.43, while features without verified capabilities averaged 1.14, a negligible gap; after fixing the language issues, the same comparison became 2.00 versus 1.07.69% similarUnverifiedThis benchmark uses 10 original short documents of 200 to 250 words each, with 18 planted anomalies, 37 scoring entries in total (18 dirty and 19 clean), testing 13 models with 74 prompts per model.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1074914API
curl https://kongchang.com/api/v1/knowledge/claims/1074914MCP
get_claim(id=1074914)